Top Countries for Business Expansion¶

Overview¶

Gapminder has collected a lot of information about how people live their lives in different countries, tracked across the years, and on a number of different indicators.

Business Goal¶

We will use metrics from this data to help stakeholders identify the best countries to open businesses targeting middle-class consumers.

Metrics to investigate:¶

  1. GDP per capita (Price and inflation-adjusted, in PPP$2021): Gross domestic product per person adjusted for differences in purchasing power (in international dollars, fixed 2021 prices, PPP based on 2021 ICP). We will use this to determine which countries have high purchasing power and market potential.
    • File: gdp_pcap_21.csv
  2. Population Size: Total population counts the number of inhabitants in the territory. We will use this to identify which countries have a big customer base.
    • File: pop.csv
  3. Population Growth (annual %): Annual population growth rate for year t is the exponential rate of growth of midyear population from year t-1 to t, expressed as a percentage. The population is based on the de facto definition of population, which counts all residents regardless of legal status or citizenship. We will use this metric to identify which countries will have a growing market.
    • File: population_growth_annual_percent.csv
  4. Urban population (% of total): Urban population refers to people living in urban areas as defined by national statistical offices. The data are collected and smoothed by United Nations Population Division. It will help determine the countries that are more likely to have better infrastructure and also urban areas are easy to target marketing and distribution.
    • File: urban_population_percent_of_total.csv
  5. Human Development Index (HDI): Human Development Index is an index used to rank countries by level of "human development". It contains three dimensions: health level, educational level, and living standard. We are going to use this to identify countries that may have skilled laborers.
    • File: hdi_human_development_index.csv
  6. Ease of doing business score (0 = lowest performance to 100 = best performance): It evaluates the regulatory environment and ease of starting and operating a business in a country.
    • File: ic_bus_dfrn_xq.csv
  7. Cost of Business Start-Up Procedures: Cost to register a business is normalized by presenting it as a percentage of gross national income (GNI) per capita. It will help identify the countries with low-start costs.
    • File: ic_reg_cost_pc_zs

We are going to use the ddf--entities--geo--country.csv file, as it contains information about each country, such as the continent they belong to.

Questions to Answer:¶

  1. Which countries had the highest GDP per capita in 2019? List the top 10.
  2. Which countries had the largest population sizes in 2019? List the top 10.
  3. Which countries had the highest population growth rates in 2019? List the top 10.
  4. Which countries had the highest Human Development Index (HDI) in 2019? List the top 10.
  5. Which countries had the best Ease of Doing Business scores in 2019? List the top 10.
  6. Which countries had the lowest costs for business start-up procedures in 2019? List the top 10.
  7. Which countries had the highest urban population rates in 2019? List the top 10.
  8. What are the relationships between the Ease of Doing Business score, the cost of business start-up procedures, and the Human Development Index by country?
  9. What are the relationships between GDP per capita, population size, and the Human Development Index by country?
  10. How did the Human Development Index (HDI) change over time for the top 10 countries in 2019?
  11. How did the Ease of Doing Business score change over time for the top 10 countries in 2019?
  12. How did the costs for business start-up procedures change over time for the 10 lowest-cost countries in 2019?
In [3]:
# import necessary libraries
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from adjustText import adjust_text  

%matplotlib inline
sns.set_style('darkgrid')
%config InlineBackend.figure_format = 'retina'

Data Gathering¶

In [5]:
df_list = {}
df_list['GDP_per_capita'] = pd.read_csv('gdp_pcap_21.csv')
df_list['population_size'] = pd.read_csv('pop.csv')
df_list['pop_growth_percent'] = pd.read_csv('population_growth_annual_percent.csv')
df_list['urban_pop_percent'] = pd.read_csv('urban_population_percent_of_total.csv')
df_list['human_dev_index'] = pd.read_csv('hdi_human_development_index.csv')
df_list['ease_of_doing_business'] = pd.read_csv('ic_bus_dfrn_xq.csv')
df_list['startup_cost'] = pd.read_csv('ic_reg_cost_pc_zs.csv')
df_countries_info = pd.read_csv('ddf--entities--geo--country.csv')

Data Assessing¶

In [7]:
df_list['GDP_per_capita'].head(5)
Out[7]:
country 1800 1801 1802 1803 1804 1805 1806 1807 1808 ... 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100
0 Afghanistan 481 481 481 481 481 481 481 481 481 ... 7320 7500 7680 7870 8060 8260 8460 8670 8880 9100
1 Angola 373 374 376 378 379 381 383 385 386 ... 29.6k 30.2k 30.7k 31.3k 31.9k 32.5k 33k 33.6k 34.2k 34.8k
2 Albania 469 471 472 473 475 476 477 479 480 ... 57.5k 58.1k 58.7k 59.2k 59.8k 60.4k 60.9k 61.5k 62.1k 62.6k
3 Andorra 1370 1370 1370 1380 1380 1380 1390 1390 1390 ... 86.5k 86.8k 87k 87.3k 87.5k 87.7k 88k 88.2k 88.4k 88.6k
4 UAE 1140 1150 1150 1150 1160 1160 1170 1170 1180 ... 92.3k 92.4k 92.4k 92.4k 92.5k 92.5k 92.5k 92.6k 92.6k 92.6k

5 rows × 302 columns

In [8]:
df_list['population_size'].head(5)
Out[8]:
country 1800 1801 1802 1803 1804 1805 1806 1807 1808 ... 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100
0 Afghanistan 3.28M 3.28M 3.28M 3.28M 3.28M 3.28M 3.28M 3.28M 3.28M ... 124M 125M 126M 126M 127M 128M 128M 129M 130M 130M
1 Angola 1.57M 1.57M 1.57M 1.57M 1.57M 1.57M 1.57M 1.57M 1.57M ... 139M 140M 142M 143M 144M 145M 147M 148M 149M 150M
2 Albania 400k 402k 404k 405k 407k 409k 411k 413k 414k ... 1.34M 1.32M 1.3M 1.29M 1.27M 1.25M 1.23M 1.22M 1.2M 1.18M
3 Andorra 2650 2650 2650 2650 2650 2650 2650 2650 2650 ... 52.8k 52.1k 51.5k 50.8k 50.2k 49.6k 49k 48.4k 47.8k 47.2k
4 UAE 40.2k 40.2k 40.2k 40.2k 40.2k 40.2k 40.2k 40.2k 40.2k ... 24.1M 24.3M 24.5M 24.7M 25M 25.2M 25.4M 25.7M 25.9M 26.1M

5 rows × 302 columns

In [9]:
df_list['pop_growth_percent'].head(5)
Out[9]:
country 1961 1962 1963 1964 1965 1966 1967 1968 1969 ... 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023
0 Aruba 2.18 1.55 1.39 1.22 1.03 0.862 0.389 −0.0857 −0.237 ... 0.692 0.638 0.59 0.537 0.495 0.452 0.134 −0.045 −0.0864 −0.158
1 Afghanistan 1.93 2.01 2.08 2.14 2.22 2.25 2.29 2.35 2.38 ... 3.66 3.12 2.58 2.87 2.89 2.91 3.13 2.85 2.53 2.67
2 Angola 1.56 1.46 1.41 1.3 1.11 0.876 0.697 0.696 1.02 ... 3.68 3.62 3.59 3.55 3.46 3.4 3.27 3.17 3.1 3.03
3 Albania 3.12 3.06 2.95 2.88 2.75 2.63 2.63 2.84 2.9 ... −0.207 −0.291 −0.16 −0.092 −0.247 −0.426 −0.574 −0.927 −1.22 −1.15
4 Andorra 7.87 7.52 7.22 6.94 6.65 7 7.92 8.13 7.72 ... 0.355 0.174 1.1 1.77 1.58 1.76 1.76 1.7 0.995 0.33

5 rows × 64 columns

In [10]:
df_list['urban_pop_percent'].head(5)
Out[10]:
country 1960 1961 1962 1963 1964 1965 1966 1967 1968 ... 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023
0 Aruba 50.8 50.80 50.70 50.70 50.70 50.7 50.7 50.7 50.7 ... 43.0 43.1 43.2 43.3 43.4 43.5 43.7 43.9 44.1 44.3
1 Afghanistan 8.4 8.68 8.98 9.28 9.59 9.9 10.2 10.6 10.9 ... 24.6 24.8 25.0 25.3 25.5 25.8 26.0 26.3 26.6 26.9
2 Angola 10.4 10.80 11.20 11.60 12.10 12.5 13.0 13.4 13.9 ... 62.7 63.4 64.1 64.8 65.5 66.2 66.8 67.5 68.1 68.7
3 Albania 30.7 30.90 31.00 31.10 31.20 31.2 31.3 31.4 31.4 ... 56.4 57.4 58.4 59.4 60.3 61.2 62.1 63.0 63.8 64.6
4 Andorra 58.5 61.00 63.50 65.90 68.20 70.4 72.6 74.6 76.6 ... 88.4 88.3 88.2 88.2 88.1 88.0 87.9 87.9 87.8 87.8

5 rows × 65 columns

In [11]:
df_list['ease_of_doing_business'].head(5)
Out[11]:
country 2015 2016 2017 2018 2019
0 Afghanistan 39.3 38.9 37.1 44.2 44.1
1 Angola 37.6 37.7 39.0 41.2 41.3
2 Albania 58.1 64.2 66.8 67.0 67.7
3 UAE 76.3 77.4 79.3 81.6 80.8
4 Argentina 56.7 57.2 57.3 58.2 59.0
In [12]:
df_list['startup_cost'].head(5)
Out[12]:
country 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
0 Afghanistan NaN 72.0 75.2 67.4 84.6 59.5 30.2 26.7 25.8 22.5 14.4 15.1 19.0 19.9 82.3 6.4 6.8
1 Angola 1320.0 910.0 654.0 498.0 344.0 197.0 151.0 227.0 163.0 143.0 130.0 119.0 17.0 20.7 17.4 13.9 11.1
2 Albania 57.1 32.3 31.3 22.5 46.1 42.5 32.2 31.8 29.3 22.4 21.2 10.1 10.3 12.5 12.0 11.3 10.8
3 UAE 18.8 17.5 15.8 13.0 13.9 11.6 11.1 12.6 13.4 11.0 11.4 11.3 11.2 13.0 13.4 22.8 17.2
4 Argentina 13.5 17.3 15.9 15.4 13.6 12.8 16.1 20.7 17.7 15.4 23.2 17.7 11.4 10.8 10.4 5.3 5.0
In [13]:
df_countries_info.head(5)
Out[13]:
country g77_and_oecd_countries income_3groups income_groups is--country iso3166_1_alpha2 iso3166_1_alpha3 iso3166_1_numeric iso3166_2 landlocked ... name un_sdg_ldc un_sdg_region un_state unhcr_region unicef_region unicode_region_subtag west_and_rest world_4region world_6region
0 abkh others NaN NaN True NaN NaN NaN NaN NaN ... Abkhazia NaN NaN False NaN NaN NaN NaN europe europe_central_asia
1 abw others high_income high_income True AW ABW 533.0 NaN coastline ... Aruba un_not_least_developed un_latin_america_and_the_caribbean False unhcr_americas NaN AW NaN americas america
2 afg g77 low_income low_income True AF AFG 4.0 NaN landlocked ... Afghanistan un_least_developed un_central_and_southern_asia True unhcr_asia_pacific sa AF rest asia south_asia
3 ago g77 middle_income lower_middle_income True AO AGO 24.0 NaN coastline ... Angola un_least_developed un_sub_saharan_africa True unhcr_southern_africa ssa AO rest africa sub_saharan_africa
4 aia others NaN NaN True AI AIA 660.0 NaN coastline ... Anguilla un_not_least_developed un_latin_america_and_the_caribbean False unhcr_americas NaN AI NaN americas america

5 rows × 23 columns

In [14]:
# check for missing values
print("Number of missing values in each dataset")
for df_name in df_list.keys():
    num_missing_values = df_list[df_name].isnull().sum().sum()
    
    print(f'{df_name}: counts: {num_missing_values}, props: {(num_missing_values/df_list[df_name].size):.3%}')

Number of missing values in each dataset
GDP_per_capita: counts: 320, props: 0.543%
population_size: counts: 100, props: 0.168%
pop_growth_percent: counts: 32, props: 0.230%
urban_pop_percent: counts: 0, props: 0.000%
human_dev_index: counts: 541, props: 8.583%
ease_of_doing_business: counts: 3, props: 0.263%
startup_cost: counts: 272, props: 7.953%
In [15]:
df_list['human_dev_index'][df_list['human_dev_index'].isnull().any(axis=1)]
Out[15]:
country 1990 1991 1992 1993 1994 1995 1996 1997 1998 ... 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
1 Angola NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.541 0.552 0.563 0.582 0.596 0.597 0.595 0.595 0.590 0.586
3 Andorra NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.869 0.864 0.871 0.867 0.871 0.868 0.872 0.873 0.848 0.858
7 Antigua and Barbuda NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.787 0.787 0.789 0.791 0.794 0.795 0.798 0.800 0.788 0.788
10 Azerbaijan NaN NaN NaN NaN NaN 0.590 0.590 0.594 0.604 ... 0.734 0.741 0.745 0.748 0.750 0.753 0.757 0.761 0.730 0.745
14 Burkina Faso NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.395 0.402 0.408 0.418 0.427 0.438 0.449 0.452 0.449 0.449
18 Bahamas NaN NaN NaN NaN NaN 0.781 0.787 0.784 0.785 ... 0.815 0.816 0.820 0.820 0.823 0.825 0.827 0.816 0.815 0.812
19 Bosnia and Herzegovina NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.745 0.751 0.756 0.761 0.770 0.772 0.776 0.783 0.781 0.780
20 Belarus NaN NaN NaN NaN NaN 0.679 0.686 0.692 0.697 ... 0.806 0.808 0.812 0.812 0.813 0.817 0.818 0.817 0.807 0.808
26 Bhutan NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.598 0.606 0.617 0.627 0.638 0.647 0.658 0.671 0.668 0.666
38 Comoros NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.533 0.539 0.540 0.544 0.548 0.553 0.557 0.560 0.562 0.558
39 Cape Verde NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.656 0.666 0.667 0.663 0.670 0.675 0.673 0.676 0.662 0.662
45 Djibouti NaN NaN NaN NaN NaN 0.351 0.350 0.353 0.355 ... 0.478 0.483 0.488 0.493 0.496 0.499 0.506 0.512 0.510 0.509
46 Dominica NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.706 0.699 0.705 0.700 0.709 0.692 0.726 0.729 0.722 0.720
52 Eritrea NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.486 0.483 0.502 0.483 0.488 0.484 0.493 0.495 0.494 0.492
55 Ethiopia NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.430 0.441 0.450 0.460 0.470 0.480 0.489 0.498 0.498 0.498
59 Micronesia, Fed. Sts. NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.628 0.626 0.625 0.626 0.629 0.632 0.633 0.633 0.629 0.628
62 Georgia NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.773 0.777 0.784 0.790 0.794 0.800 0.804 0.810 0.802 0.802
66 Guinea-Bissau NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.456 0.460 0.466 0.472 0.478 0.481 0.482 0.490 0.483 0.483
67 Equatorial Guinea NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.592 0.594 0.598 0.603 0.603 0.601 0.601 0.605 0.599 0.596
69 Grenada NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.784 0.787 0.787 0.790 0.791 0.794 0.797 0.800 0.792 0.795
74 Croatia NaN NaN NaN NaN NaN 0.712 0.724 0.732 0.742 ... 0.830 0.838 0.841 0.843 0.848 0.852 0.856 0.861 0.855 0.858
92 Kiribati NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.594 0.602 0.610 0.622 0.617 0.621 0.622 0.630 0.623 0.624
93 St. Kitts and Nevis NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.766 0.770 0.772 0.772 0.779 0.778 0.779 0.783 0.779 0.777
97 Lebanon NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.762 0.759 0.750 0.746 0.743 0.750 0.750 0.745 0.726 0.706
98 Liberia NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.470 0.475 0.472 0.473 0.478 0.481 0.483 0.484 0.480 0.481
101 Liechtenstein NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.923 0.922 0.920 0.924 0.923 0.933 0.928 0.940 0.933 0.935
109 Madagascar NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.496 0.499 0.502 0.504 0.505 0.507 0.507 0.510 0.501 0.501
110 Maldives NaN NaN NaN NaN NaN 0.582 0.593 0.600 0.606 ... 0.709 0.718 0.728 0.736 0.741 0.746 0.750 0.755 0.734 0.747
112 Marshall Islands NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.628 0.632 0.633 0.638 0.637 0.638 0.639 0.640 0.639 0.639
113 North Macedonia NaN NaN NaN NaN NaN 0.642 0.650 0.655 0.661 ... 0.743 0.750 0.755 0.762 0.767 0.773 0.779 0.784 0.774 0.770
117 Montenegro NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.812 0.817 0.819 0.822 0.824 0.830 0.834 0.837 0.826 0.832
126 Nigeria NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.499 0.506 0.512 0.516 0.521 0.526 0.531 0.538 0.535 0.535
132 Oman NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.802 0.809 0.814 0.823 0.833 0.831 0.834 0.839 0.827 0.816
137 Palau NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.776 0.779 0.779 0.780 0.778 0.775 0.778 0.776 0.773 0.767
142 Palestine NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.701 0.703 0.698 0.710 0.715 0.719 0.723 0.727 0.716 0.715
146 Rwanda NaN NaN NaN NaN NaN 0.286 0.296 0.302 0.309 ... 0.506 0.508 0.513 0.515 0.524 0.526 0.528 0.534 0.532 0.534
151 Solomon Islands NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.554 0.559 0.558 0.559 0.560 0.564 0.566 0.567 0.565 0.564
154 San Marino NaN NaN NaN NaN NaN NaN NaN NaN NaN ... NaN NaN NaN NaN NaN NaN 0.860 0.862 0.845 0.853
155 Serbia NaN NaN NaN NaN NaN 0.674 0.679 0.685 0.687 ... 0.780 0.785 0.788 0.794 0.800 0.802 0.808 0.811 0.804 0.802
156 South Sudan NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.397 0.411 0.410 0.412 0.402 0.395 0.395 0.393 0.386 0.385
158 Suriname NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.731 0.734 0.736 0.744 0.747 0.750 0.755 0.755 0.743 0.730
160 Slovenia NaN NaN NaN NaN NaN 0.760 0.771 0.782 0.794 ... 0.891 0.900 0.902 0.903 0.910 0.913 0.917 0.921 0.913 0.918
163 Seychelles NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.786 0.787 0.796 0.796 0.796 0.796 0.800 0.802 0.793 0.785
165 Chad NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.381 0.386 0.393 0.389 0.391 0.393 0.398 0.403 0.397 0.394
169 Turkmenistan NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.727 0.730 0.734 0.740 0.742 0.741 0.746 0.742 0.741 0.745
170 Timor-Leste NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.635 0.630 0.622 0.614 0.604 0.606 0.605 0.614 0.614 0.607
181 Uzbekistan NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.687 0.693 0.698 0.701 0.709 0.715 0.720 0.726 0.721 0.727
182 St. Vincent and the Grenadines NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.737 0.743 0.752 0.759 0.765 0.774 0.775 0.769 0.764 0.751
185 Vanuatu NaN NaN NaN NaN NaN NaN NaN NaN NaN ... 0.590 0.593 0.595 0.595 0.596 0.599 0.603 0.611 0.608 0.607
186 Samoa NaN NaN NaN NaN NaN 0.659 0.668 0.670 0.674 ... 0.709 0.710 0.711 0.716 0.717 0.716 0.716 0.715 0.712 0.707

50 rows × 33 columns

Most of the missing values in the Human Development Index data are from earlier years.

Quality Issues¶

  • Inconsistent time frames across datasets. We will use the time frame spanning from 2015 to 2019.
  • There are missing values
  • There are 'K' and 'M' used to describe thousands and millions in GDP_per_capita and population_size
  • Incorrect datatypes

Tidiness Issues¶

  • Year values are used as column names instead of a single "Year" variable.
  • The data need to be combined into a single dataframe with columns for country, year, region, indicator 1 value, indicator 2 value, and so on.

Cleaning data¶

Inconsistent time frames across datasets.¶

Define¶
  • We will analyze the data using the common time frame across datasets, which spans from 2015 to 2019, as indicated by the time frame in the ease_of_doing_business dataset.
Code¶
In [21]:
print(df_list.keys())

dict_keys(['GDP_per_capita', 'population_size', 'pop_growth_percent', 'urban_pop_percent', 'human_dev_index', 'ease_of_doing_business', 'startup_cost'])
In [22]:
# Extract the common column names from the ease_of_doing_business dataset and use them to standardize the columns across all datasets.
common_cols = df_list['ease_of_doing_business'].columns
print(common_cols)

Index(['country', '2015', '2016', '2017', '2018', '2019'], dtype='object')
In [23]:
for df_name in df_list.keys():
    if df_name != 'ease_of_doing_business':
        df_list[df_name] = df_list[df_name][common_cols]
Test¶

Ensure that all datasets contain only the following columns: country, 2015, 2016, 2017, 2018, and 2019.

In [25]:
for df_name in df_list.keys():
    if (df_list[df_name].columns).all() != common_cols.all():
        print(df_name)
In [26]:
df_list['GDP_per_capita'].head()
Out[26]:
country 2015 2016 2017 2018 2019
0 Afghanistan 2970 2960 2960 2910 2930
1 Angola 9990 9390 9050 8630 8280
2 Albania 13.2k 13.7k 14.2k 14.8k 15.2k
3 Andorra 62.5k 64.1k 63.2k 63.2k 63.3k
4 UAE 67k 70.1k 70k 70.4k 70.6k

Missing data¶

First, review the missing values again. You will notice that the number of missing values has decreased after selecting the time interval for our analysis.

In [29]:
# check for missing values
print("Number of missing values in each dataset")
for df_name in df_list.keys():
    num_missing_values = df_list[df_name].isnull().sum().sum()
    
    print(f'{df_name}: counts: {num_missing_values}, props: {(num_missing_values/df_list[df_name].size):.3%}')

Number of missing values in each dataset
GDP_per_capita: counts: 10, props: 0.855%
population_size: counts: 5, props: 0.423%
pop_growth_percent: counts: 0, props: 0.000%
urban_pop_percent: counts: 0, props: 0.000%
human_dev_index: counts: 3, props: 0.262%
ease_of_doing_business: counts: 3, props: 0.263%
startup_cost: counts: 3, props: 0.263%
Define¶

Drop missing values, as most of them fall outside the time frame used for our analysis.

Code¶
In [32]:
for df_name in df_list.keys():
    if df_list[df_name].isnull().sum().sum() != 0:
        df_list[df_name].dropna(inplace=True, axis=0)
Test¶
In [34]:
for df_name in df_list.keys():
    print(f'{df_name}: ', df_list[df_name].isnull().sum().sum())

GDP_per_capita:  0
population_size:  0
pop_growth_percent:  0
urban_pop_percent:  0
human_dev_index:  0
ease_of_doing_business:  0
startup_cost:  0

Year values are used as column names instead of a single "Year" variable.¶

Define¶
  • Use the pandas.melt() function to create a new column for years, converting the year columns (2015, 2016, 2017, etc.) into a single column named year.
Code¶
In [38]:
for df_name in df_list.keys():
    df_list[df_name] = pd.melt(df_list[df_name], id_vars='country',
                               value_vars=['2015', '2016', '2017', '2018', '2019'],
                               value_name=df_name, var_name='year')
Test¶
In [40]:
# ensure that each dataset has only 3 columns; country, year, and the metric value
for df_name in df_list.keys():
    print(f'{df_name}: Shape: {df_list[df_name].shape}, Columns: ', df_list[df_name].columns.tolist())

GDP_per_capita: Shape: (965, 3), Columns:  ['country', 'year', 'GDP_per_capita']
population_size: Shape: (980, 3), Columns:  ['country', 'year', 'population_size']
pop_growth_percent: Shape: (1085, 3), Columns:  ['country', 'year', 'pop_growth_percent']
urban_pop_percent: Shape: (1075, 3), Columns:  ['country', 'year', 'urban_pop_percent']
human_dev_index: Shape: (950, 3), Columns:  ['country', 'year', 'human_dev_index']
ease_of_doing_business: Shape: (945, 3), Columns:  ['country', 'year', 'ease_of_doing_business']
startup_cost: Shape: (945, 3), Columns:  ['country', 'year', 'startup_cost']
In [41]:
df_list['GDP_per_capita'].head(2)
Out[41]:
country year GDP_per_capita
0 Afghanistan 2015 2970
1 Angola 2015 9990

The data need to be combined into a single dataframe with columns for country, year, region, indicator 1 value, indicator 2 value, and so on.¶

Define¶
  • Join the data into a single dataset using pandas' merge function.
Code¶

Use the pop_growth_percent dataset to perform a left join, as it has the largest number of records. This ensures we don't lose important information until we investigate the countries with missing values.

In [45]:
df = df_list['pop_growth_percent']
for df_name in df_list.keys():
    if df_name != 'pop_growth_percent':
        df = df.merge(df_list[df_name], on=['country', 'year'], how='left')
Test¶
In [47]:
df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 1085 entries, 0 to 1084
Data columns (total 9 columns):
 #   Column                  Non-Null Count  Dtype  
---  ------                  --------------  -----  
 0   country                 1085 non-null   object 
 1   year                    1085 non-null   object 
 2   pop_growth_percent      1085 non-null   object 
 3   GDP_per_capita          960 non-null    object 
 4   population_size         975 non-null    object 
 5   urban_pop_percent       1075 non-null   float64
 6   human_dev_index         950 non-null    float64
 7   ease_of_doing_business  945 non-null    float64
 8   startup_cost            945 non-null    float64
dtypes: float64(4), object(5)
memory usage: 76.4+ KB
In [48]:
df.isnull().sum()
Out[48]:
country                     0
year                        0
pop_growth_percent          0
GDP_per_capita            125
population_size           110
urban_pop_percent          10
human_dev_index           135
ease_of_doing_business    140
startup_cost              140
dtype: int64
In [49]:
df.year.value_counts()
Out[49]:
year
2015    217
2016    217
2017    217
2018    217
2019    217
Name: count, dtype: int64
  • The new dataset contains 1,085 entries and 9 columns.
  • Missing values are present.
  • The datatype of the year column needs to be converted to an integer.
  • As mentioned earlier, 'K' and 'M' are used to represent thousands and millions in the GDP_per_capita and population_size columns.
  • Additionally, the pop_growth_percent column requires further investigation, and its datatype should be converted to float.
In [51]:
print(df[df.isnull().any(axis=1)].country.unique())

['Aruba' 'Andorra' 'American Samoa' 'Bermuda' 'Channel Islands' 'Cuba'
 'Cayman Islands' 'Eritrea' 'Faeroe Islands' 'Isle of Man' 'Gibraltar'
 'Greenland' 'Guam' 'Kosovo' 'Liechtenstein' 'Macao, China'
 'St. Martin (French part)' 'Monaco' 'Northern Mariana Islands'
 'New Caledonia' 'Curaçao' 'Nauru' 'Puerto Rico' 'North Korea'
 'French Polynesia' 'San Marino' 'Somalia' 'South Sudan'
 'Sint Maarten (Dutch part)' 'Turks and Caicos Islands' 'Turkmenistan'
 'Tuvalu' 'British Virgin Islands' 'Virgin Islands (U.S.)']

Many of the countries in this list are either small territories, island nations, or countries with limited economic data, which may not be relevant for stakeholders looking to invest in larger, more stable markets. We will drop those countries

Missing values are present.¶

Define¶

Drop rows that have missing values

Code¶
In [55]:
countries_to_drop = df[df.isnull().any(axis=1)].country.values
df = df[~df['country'].isin(countries_to_drop)]

Test¶

In [57]:
df.isnull().sum()
Out[57]:
country                   0
year                      0
pop_growth_percent        0
GDP_per_capita            0
population_size           0
urban_pop_percent         0
human_dev_index           0
ease_of_doing_business    0
startup_cost              0
dtype: int64

The datatype of the year column needs to be converted to an integer.¶

As mentioned earlier, 'K' and 'M' are used to represent thousands and millions in the GDP_per_capita and population_size columns.¶

Additionally, the pop_growth_percent column requires further investigation, and its datatype should be converted to float.¶

Define¶
  • Use pandas.to_numeric function to convert the datatype of the year column to int
  • Replace the observations containing 'K' and 'M' in the GDP_per_capita and population_size columns with their corresponding numeric values, and convert their datatype to float.
  • Convert the datatype of the pop_growth_percent column to float.

Code¶

In [61]:
df.year = pd.to_numeric(df.year)
In [62]:
def get_numeric_val(strNum):
    if 'k' in strNum:
        return float(strNum.replace('k', '')) * 1_000
    elif 'M' in strNum:
        return float(strNum.replace('M', '')) * 1_000_000
    elif 'B' in strNum:
        return float(strNum.replace('B', '')) * 1_000_000_000
    else:
        return float(strNum)
In [63]:
df.GDP_per_capita = df.GDP_per_capita.apply(get_numeric_val)
df.population_size = df.population_size.apply(get_numeric_val)
In [64]:
df.pop_growth_percent = df.pop_growth_percent.str.replace('−', '-').astype(float)
Test¶
In [66]:
df.dtypes
Out[66]:
country                    object
year                        int64
pop_growth_percent        float64
GDP_per_capita            float64
population_size           float64
urban_pop_percent         float64
human_dev_index           float64
ease_of_doing_business    float64
startup_cost              float64
dtype: object

We want to add country region from the countries dataset to our dataset.¶

Define¶

Use pandas.merge to add country region from the countries dataset to our dataset.

Code¶
In [69]:
df_countries_info.head(1)
Out[69]:
country g77_and_oecd_countries income_3groups income_groups is--country iso3166_1_alpha2 iso3166_1_alpha3 iso3166_1_numeric iso3166_2 landlocked ... name un_sdg_ldc un_sdg_region un_state unhcr_region unicef_region unicode_region_subtag west_and_rest world_4region world_6region
0 abkh others NaN NaN True NaN NaN NaN NaN NaN ... Abkhazia NaN NaN False NaN NaN NaN NaN europe europe_central_asia

1 rows × 23 columns

In [70]:
df = pd.merge(df, df_countries_info[['name', 'world_6region']], how='left', left_on='country', right_on='name')
df.drop('name', axis=1, inplace=True)
Test¶
In [72]:
df.head()
Out[72]:
country year pop_growth_percent GDP_per_capita population_size urban_pop_percent human_dev_index ease_of_doing_business startup_cost world_6region
0 Afghanistan 2015 3.120 2970.0 33800000.0 24.8 0.478 39.3 19.0 south_asia
1 Angola 2015 3.620 9990.0 28200000.0 63.4 0.582 37.6 17.0 sub_saharan_africa
2 Albania 2015 -0.291 13200.0 2900000.0 57.4 0.795 58.1 10.3 europe_central_asia
3 UAE 2015 0.912 67000.0 8670000.0 85.7 0.865 76.3 11.2 middle_east_north_africa
4 Argentina 2015 1.080 28900.0 43500000.0 91.5 0.848 56.7 11.4 america
In [73]:
# ensure that we don't have missing values after mergin the region column
df.isnull().sum().sum()
Out[73]:
0
In [74]:
df.world_6region = df.world_6region.str.replace('_', ' ').str.title()

Eventually, our data is ready for analysis.

Exploratory Data Analysis¶

In [77]:
df.describe()
Out[77]:
year pop_growth_percent GDP_per_capita population_size urban_pop_percent human_dev_index ease_of_doing_business startup_cost
count 915.000000 915.000000 915.000000 9.150000e+02 915.000000 915.000000 915.000000 915.000000
mean 2017.000000 1.299604 23009.077596 4.117335e+07 58.344372 0.720609 61.796721 23.421967
std 1.414987 1.313483 23719.664816 1.490484e+08 22.805022 0.149633 13.272941 37.959452
min 2015.000000 -4.420000 234.000000 1.780000e+04 12.100000 0.376000 30.000000 0.000000
25% 2016.000000 0.418500 5470.000000 2.210000e+06 40.200000 0.600000 52.850000 2.300000
50% 2017.000000 1.220000 14900.000000 8.870000e+06 58.300000 0.744000 61.300000 10.400000
75% 2018.000000 2.230000 32700.000000 3.050000e+07 77.400000 0.835500 72.700000 26.800000
max 2019.000000 9.220000 135000.000000 1.420000e+09 100.000000 0.962000 87.200000 393.000000
  • GDP per capita exceeds $100,000 in some countries. These extreme values skew the mean to be higher than the median.
  • Some countries have a negative population growth percentage, indicating a declining population.
  • Certain countries have an urban population of 100%.
  • Approximately 25% of countries have a Human Development Index (HDI) higher than 0.84.
  • About 25% of countries have an Ease of Doing Business Index score higher than 72.7.
  • Some countries have a Cost of Business Start-Up Procedures equivalent to 400% of their Gross National Income (GNI) per capita.
In [79]:
df.hist(figsize=(10, 10));
plt.show()
No description has been provided for this image
  • The distribution of Gross Domestic Product (GDP) per capita is right-skewed.
  • The distribution of the Cost of Business Start-Up Procedures is also right-skewed.
  • The distribution of the population growth percentage follows a normal distribution.
  • The distributions of the Ease of Doing Business Index and the Human Development Index are slightly left-skewed.
In [81]:
pd.plotting.scatter_matrix(df, figsize=(15, 15));
plt.show()
No description has been provided for this image
  • The Human Development Index (HDI) and the Ease of Doing Business Index show a positive correlation.
  • The Human Development Index (HDI) and the Urban Population Percentage also demonstrate a positive correlation.
  • The Cost of Business Start-Up Procedures and the Ease of Doing Business Index exhibit a negative correlation.

Since we are going to answer questions about the data for the most recent year, which is 2019, let's filter the dataset accordingly.

In [84]:
df_19 = df[df.year == 2019]
print(df_19.year.value_counts())
df_19 = df_19.drop('year', axis=1)

year
2019    183
Name: count, dtype: int64

Draw Conclustions¶

Let's answer our questions about the data to extract insights and trends that will help stakeholders make informed decisions about the countries they should invest in.

In [86]:
title_fontdict = {
    'fontsize': 18,           # Font size
    'fontweight': 'bold',     # Bold font
    'fontstyle': 'normal',    # Normal style (can be 'italic' or 'oblique')
    'fontfamily': 'sans-serif',  # Font family (e.g., 'serif', 'sans-serif', 'monospace')
    'color': 'black',         # Text color
}

label_fontdict = {
    'fontsize': 14,           # Font size
    'fontweight': 'normal',   # Normal weight (can be 'bold', 'light', etc.)
    'fontstyle': 'normal',    # Normal style (can be 'italic' or 'oblique')
    'fontfamily': 'sans-serif',  # Font family (e.g., 'serif', 'sans-serif', 'monospace')
}
In [87]:
def format_tick_labels(nums):
    """Formats numbers into human-readable strings with metric suffixes.

    Converts numbers to strings with appropriate metric suffixes (k, M, B, T),
    preserving decimal precision where needed. Handles negative numbers and
    includes proper rounding.

    Args:
        nums (list[int/float]): List of numbers to convert.

    Returns:
        list[str]: List of formatted strings using metric suffixes:
            - < 1000: no suffix (e.g., '867')
            - >= 1000: k (e.g., '1.2k')
            - >= 1M: M (e.g., '3.4M')
            - >= 1B: B (e.g., '2.5B')
            - >= 1T: T (e.g., '1.2T')

    Example:
        >>> format_tick_labels([867, 1234, 2500000, 3500000000, -1234567])
        ['867', '1.2k', '2.5M', '3.5B', '-1.2M']
    """
    suffixes = ['', 'k', 'M', 'B', 'T']
    
    def _format_number(n):
        n = float(n)
        magnitude = 0
        abs_n = abs(n)
        
        # Find the appropriate magnitude
        while abs_n >= 1000 and magnitude < len(suffixes) - 1:
            magnitude += 1
            abs_n /= 1000.0
            
        # Round to 1 decimal place if there's a decimal component
        formatted = f'{abs_n:.1f}'.rstrip('0').rstrip('.')
        
        # Add negative sign if needed
        if n < 0:
            formatted = f'-{formatted}'
            
        return f'{formatted}{suffixes[magnitude]}'
    
    return [_format_number(n) for n in nums]

def plot_indicator_by_country(df, indicator, title, xlabel=None, ylabel=None, format_xticks=False, top_result=True):
    """Creates a horizontal bar plot showing the top 10 countries for a given indicator.

    Takes a DataFrame with country and indicator data, and creates a horizontal bar plot 
    showing the top 10 countries with the highest values for the specified indicator.
    The function automatically sorts the data and displays the result.

    Args:
        df (pandas.DataFrame): DataFrame containing at least 'country' column and the indicator column.
        indicator (str): Name of the column in df to plot.
        title (str): Title of the plot.
        xlabel (str, optional): Label for x-axis. Defaults to None.
        ylabel (str, optional): Label for y-axis. Defaults to None.
        xticks (bool, optional): Whether to format x-axis tick labels. Defaults to False.

    Returns:
        None: Displays the plot using plt.show()

    Example:
        >>> df = pd.DataFrame({
        ...     'country': ['USA', 'China', 'Japan'],
        ...     'gdp': [20.94, 14.72, 5.06]
        ... })
        >>> plot_indicator_by_country(df, 'gdp', 'GDP by Country', 'GDP (Trillion USD)', 'Country', xticks=True)
    """
    results = None
    if top_result:
        # Get top 10 countries
        results = df.groupby('country')[indicator].first().sort_values(ascending=True)[-10:]
    else:
        results = df.groupby('country')[indicator].first().sort_values(ascending=True)[:10]
    # Create positions for the bars
    positions = np.arange(len(results))
    
    # Create the plot
    plt.figure(figsize=(10, 6))
    plt.barh(positions, results.values)
    
    # Set country names on y-axis
    plt.yticks(positions, results.index)
    
    # Add labels and title
    if title:
        plt.title(title, fontdict=title_fontdict, pad=20)
    if xlabel:
        plt.xlabel(xlabel, fontdict=label_fontdict, labelpad=10)
    if ylabel:
        plt.ylabel(ylabel, fontdict=label_fontdict, labelpad=10)
        
    # Format x-axis ticks if requested
    if format_xticks:
        current_ticks = plt.xticks()[0]  # Get current tick positions
        plt.xticks(current_ticks, format_tick_labels(current_ticks))
    
    # Add spacing and ensure layout looks good
    plt.margins(x=0.1)
    plt.tight_layout()
    
    plt.show()
    return results.index.tolist()[-10:] if top_result else results.index.tolist()[:10]

1. Which countries had the highest GDP per capita in 2019? List the top 10.¶

In [89]:
top_countries = []
top_countries += plot_indicator_by_country(df_19, indicator='GDP_per_capita', 
                          title= 'Which countries had the highest GDP per capita in 2019?',
                          xlabel= 'Gross Domestic Product per Capita', format_xticks=True)
No description has been provided for this image
  • In 2019, Luxembourge has the highest domestic gross product per captia which imply that its citizens may have high purchasing power and market potential.

2. Which countries had the largest population sizes in 2019? List the top 10.¶

In [92]:
top_countries += plot_indicator_by_country(df_19, indicator='population_size', 
                          title='Which countries had the largest popualtion size in 2019?', 
                          xlabel='Population Size',
                         format_xticks=True)
No description has been provided for this image
  • China and India have significantly larger populations compared to other countries, which indicates a substantial potential customer base for businesses.
  • The USA is one of the top 10 countries with both a high GDP per capita and a large population size, making it a potentially lucrative market with strong purchasing power and a substantial customer base.

3. Which countries had the highest population growth rates in 2019? List the top 10.¶

In [95]:
top_countries += plot_indicator_by_country(df_19, indicator='pop_growth_percent', 
                          title='Which countries had the highest population growth rates in 2019?')
No description has been provided for this image
  • Malta, Syria, and Niger have the highest population growth percentages, making them strong candidates for businesses that target products for children. These countries are expected to have growing markets due to their increasing population size. Since Syria do not have political or economic stability, I don't think it will be a good canditate to invest in.

4. Which countries had the highest Human Development Index (HDI) in 2019? List the top 10.¶

In [98]:
top_countries += plot_indicator_by_country(df_19, indicator='human_dev_index', 
                          title='Which countries had the highest Human Development Index (HDI) in 2019?')
No description has been provided for this image
  • Switzerland, Norway, and Iceland have the highest Human Development Index (HDI), suggesting a higher likelihood of a skilled and educated workforce. This is crucial for businesses that require specialized skills to thrive.

5. Which countries had the best Ease of Doing Business scores in 2019? List the top 10.¶

In [101]:
top_countries += plot_indicator_by_country(df_19, indicator='ease_of_doing_business', 
                          title='Which countries had the best Ease of Doing Business scores in 2019?')
No description has been provided for this image
  • New Zealand, Singapore, and Hong Kong, China have the highest Ease of Doing Business scores, reflecting favorable regulatory environments and the ease of starting and operating a business. Additionally, some countries excel in both Ease of Doing Business and Human Development Index (HDI), indicating the presence of skilled labor and a business-friendly environment, such as Singapore and Hong Kong, China

6. Which countries had the lowest costs for business start-up procedures in 2019? List the top 10.¶

In [104]:
top_countries += plot_indicator_by_country(df_19, indicator='startup_cost', 
                          title='Which countries had the lowest costs for business start-up procedures in 2019?', 
                          top_result=False)
No description has been provided for this image
  • United Kingdom, Slovenia, and Rwanda have no costs associated with business start-up procedures, making them attractive for new businesses. Additionally, countries with high ease of doing business scores generally tend to have lower costs for business start-up procedures, creating a favorable environment for entrepreneurs.

7. Which countries had the highest urban population rates in 2019? List the top 10.¶

In [107]:
top_countries += plot_indicator_by_country(df_19, indicator='urban_pop_percent', 
                          title='Which countries had the highest urban population rates in 2019?')
No description has been provided for this image
  • Hong Kong, China, Singapore, and Kuwait have 100% urban populations, which suggests they are likely to have better infrastructure. Urban areas also offer advantages for businesses due to easier targeting of marketing and distribution efforts, making these countries attractive for investment.

8. What are the relationships between the Ease of Doing Business score, the cost of business start-up procedures, and the Human Development Index by country?¶

In [110]:
top_countries = set(top_countries) 
len(top_countries)
Out[110]:
49
In [111]:
startup_cost_median = df_19.startup_cost.median()
ease_business_median = df_19.ease_of_doing_business.median()
human_idx_median = df_19.human_dev_index.median()
In [112]:
# Create the scatter plot
plt.figure(figsize=(15, 10))
scatter = plt.scatter(df_19.query('country in @top_countries').startup_cost, 
                      df_19.query('country in @top_countries').ease_of_doing_business, 
                      alpha=0.7, c=df_19.query('country in @top_countries').human_dev_index, cmap='flare')

# Add title and labels
plt.title("Relationships Between Ease of Business, Start-Up Costs, \nand Human Development Index Across Countries", 
          fontdict=title_fontdict)
plt.xlabel('Cost of business start-up procedures', fontdict=label_fontdict)
plt.ylabel('Ease of Doing Business score', fontdict=label_fontdict)
color_bar = plt.colorbar()
color_bar.ax.set_title('HDI', fontweight='bold', fontsize=16)
# Add median lines
plt.axhline(y=ease_business_median, color='gray', linestyle='--', alpha=0.7)
plt.axvline(x=startup_cost_median, color='gray', linestyle='--', alpha=0.7)
plt.xlim([-1, 50])
plt.ylim([50, 90])
# Prepare text annotations
texts = []
for c in top_countries:
    x_val = df_19[df_19.country == c]['startup_cost'].values[0] 
    y_val = df_19[df_19.country == c]['ease_of_doing_business'].values[0]
    if x_val < 50 and y_val > 50:
        texts.append(plt.text(x=x_val + 0.4, 
                              y=y_val - 0.2, 
                              s=c, fontsize=7.5, alpha=0.7, fontweight='bold'))

# Adjust text labels to avoid overlap
adjust_text(texts, arrowprops=dict(arrowstyle='->', color='gray', lw=0.5))
plt.show()
No description has been provided for this image

New Zealand, Singapore, the USA, Hong Kong, China, Denmark, and the United Kingdom have high Ease of Doing Business scores and Human Development Index (HDI) values and low costs for business start-up procedures.

  • These countries have streamlined regulatory environments, making starting and operating businesses easier.
  • They also have well-educated and healthy workforces, making them ideal for industries that require skilled labor.
  • Additionally, They have low financial barriers to entry, reducing the initial investment required to establish a business.

To narrow down the scope of target countries for potential investment, we will focus on countries meeting the following criteria:

  1. Ease of Doing Business Score: Greater than the median.
  2. Human Development Index (HDI): Greater than the median.
  3. Cost of Start-Up Procedures: Below the median.

This approach will help identify countries with favorable business environments, skilled workforces, and minimal financial barriers to entry. Let's analyze the data and keep track of these countries.

In [114]:
startup_cost_median = df_19.startup_cost.median()
ease_business_median = df_19.ease_of_doing_business.median()
human_idx_median = df_19.human_dev_index.median()

top_countries = df_19.query(
    'startup_cost < @startup_cost_median and ease_of_doing_business > @ease_business_median '
    'and human_dev_index > @human_idx_median and country in @top_countries' 
).country.values.tolist()

9. What are the relationships between GDP per capita, population size, and the Human Development Index by country?¶

In [116]:
top_countries.remove('China')
In [117]:
# Create the scatter plot
plt.figure(figsize=(15, 10))
scatter = plt.scatter(df_19.query('country in @top_countries').population_size, 
                      df_19.query('country in @top_countries').GDP_per_capita, 
                      alpha=0.7, c=df_19.query('country in @top_countries').human_dev_index, cmap='flare')

# Add title and labels
plt.title("Relationships Between GDP per Capita, Population Size, \nand Human Development Index Across Countries", 
          fontdict=title_fontdict)
plt.xlabel('Population Size', fontdict=label_fontdict)
plt.ylabel('Gross Demostic Product per Capita', fontdict=label_fontdict)

color_bar = plt.colorbar()
color_bar.ax.set_title('HDI', fontweight='bold', fontsize=16)
current_ticks = plt.xticks()[0]
plt.xticks(current_ticks, labels=format_tick_labels(current_ticks))
# Prepare text annotations
texts = []
for c in top_countries:
    x_val = df_19[df_19.country == c]['population_size'].values[0] 
    y_val = df_19[df_19.country == c]['GDP_per_capita'].values[0]
 
    texts.append(plt.text(x=x_val + 3_000_000, 
                            y=y_val - 1000, 
                            s=c, fontsize=7.5, alpha=0.7, fontweight='bold'))

# Adjust text labels to avoid overlap
adjust_text(texts, arrowprops=dict(arrowstyle='->', color='gray', lw=0.5))
plt.show()
No description has been provided for this image

  • China has a very high population size but a relatively low Human Development Index (HDI) and GDP per capita compared to other countries. In contrast, Hong Kong, China, has a higher GDP per capita and HDI, indicating better economic and developmental conditions.

  • I will exclude China from my top countries list to focus on the population sizes of other countries in the chart above.

  • The USA has a large population size compared to other countries, and it also has a strong GDP per capita and Human Development Index (HDI). This combination makes the USA an ideal candidate for businesses targeting a large customer base, consumers with disposable income and high purchasing power, and industries that require skilled labor.

  • Japan, Russia, the UK, and Germany come next, with populations exceeding 50 million. They also have good GDP per capita and HDI, making them solid candidates for businesses targeting skilled labor, high purchasing power, and moderate customer base with market potential.

  • Luxembourg, Singapore, and Qatar, despite their small population sizes, boast high GDP per capita and HDI, making them great candidates for businesses targeting high purchasing power and skilled labor.

In [119]:
pop_size_median = df_19.population_size.median()
gdp_median = df_19.GDP_per_capita.median()
top_countries = df_19.query(
                'population_size > @pop_size_median and GDP_per_capita > @gdp_median '
                'and human_dev_index > @human_idx_median and country in @top_countries' 
            ).country.values.tolist()

10. How did the Human Development Index (HDI) change over time for the top 10 countries in 2019?¶

In [121]:
import matplotlib.pyplot as plt
import seaborn as sns
import matplotlib.colors as mcolors

def plot_indicator_trends(df, indicator, title, xlabel):
    # Extended colorblind-friendly palette with 11 colors
    # Combining colors from colorblind-safe palettes and IBM's color-blind safe palette
    colors = [
        '#0173B2',  # Blue
        '#DE8F05',  # Orange
        '#029E73',  # Green
        '#D55E00',  # Red
        '#CC78BC',  # Purple
        '#CA9161',  # Brown
        '#FBAFE4',  # Pink
        '#949494',  # Gray
        '#ECE133',  # Yellow
        '#56B4E9',  # Light Blue
        '#004F92'   # Dark Blue
    ]
    
    plt.figure(figsize=(10, 8))
    
    for i, c in enumerate(top_countries):
        plt.plot(df.query('country == @c').year, 
                df.query('country == @c')[indicator], 
                marker='o', 
                linestyle='-', 
                label=c,
                color=colors[i])
    
    plt.title(title, fontdict=title_fontdict)
    plt.ylabel(xlabel, fontdict=label_fontdict)
    plt.legend(loc='best')
    plt.xticks([2015, 2016, 2017, 2018, 2019])
    plt.grid(True, alpha=0.3)
    plt.show()
In [122]:
plot_indicator_trends(df, 'human_dev_index', 'Trends in Human Development Index Over Time for the Top 10 Countries',
                    'Human Development Index')
No description has been provided for this image
  • In general, the Human Development Index (HDI) has been increasing over time for the top 10 countries, indicating overall improvements in education, healthcare, and income levels.
  • For the United Kingdom, the HDI slightly decreased in 2018 before increasing again in 2019, showing some fluctuations in its human development indicators.
  • All these countries are excellent candidates for businesses that require skilled labor, as they have well-educated and healthy workforces, making them ideal markets for industries that depend on specialized talent.

11. How did the Ease of Doing Business score change over time for the top 10 countries in 2019?¶

In [125]:
plot_indicator_trends(df, 'ease_of_doing_business', 'Trends in Ease of Doing Business Over Time for the Top 10 Countries',
                    'Ease of Doing Business Scores')
No description has been provided for this image
  • Ease of Doing Business scores have remained stable over time for countries such as the USA, the United Kingdom, Sweden, Canada, and Germany, indicating consistent regulatory environments that support business operations.
  • On the other hand, Ease of Doing Business scores have been increasing over time for countries such as Kazakhstan, Russia, and Belgium, suggesting improvements in their regulatory frameworks and business environments.

12. How did the costs for business start-up procedures change over time for the 10 lowest-cost countries in 2019?¶

In [128]:
plot_indicator_trends(df, 'startup_cost', 'Trends in Costs of Start-up Procedures Over Time for the Top 10 Countries',
                    'Costs of Start-up Procedures')
No description has been provided for this image
  • The costs of start-up procedures are very low for countries such as the United Kingdom, Kazakhstan, Canada, and Sweden, making these countries favorable for new business ventures.
  • In Romania, the costs of start-up procedures decreased in 2017 and then stabilized, indicating a positive shift in the business environment during that period.
  • Overall, while there are slight differences in the cost structures, all these countries are generally suitable for starting a business, depending on the specific business type, location, and other factors.